1

North East Indian Linguistics Society

http://www.latrobe.edu.au/rclt/Neils/Neils.htm

5th International Conference of the North East Indian Linguistics Society Guwahati, Assam, India, February 12-14, 2010

Abstracts

(alphabetically by surname of (first) author)

Bugun language and community Madhumita Barbora

Tezpur University Khowa (Khova) or Bugun belongs to the Kho-Bwa cluster of languages. According to van Driem (2007) the four languages of the Kho-Bwa cluster in western Arunachal Pradesh, just east of Bhutan, are “all endangered with imminent extinction”. Bugun speakers themselves fear that their mother tongue may become extinct in the near future.

The Buguns inhabit in the southern part of West Kameng district. They live in 10 villages and hamlets in Bomdila Circle. The names of the villages/hamlets are: Wanghoo, Sinchung, New Kaspi, Ramo, Namfri, Chithu, Sachida, Pani-Phu and Diching, as well as a small number of additional locales. These villages/hamlets are located on mountain slopes on both sides of the Rupa River, within a distance of 70–80 kilometres around Bomdila Circle. Accessibility to some of these villages is difficult because of the steep mountains. Some of the villages are accessible only by foot. The Bugun area is surrounded on the North by the Monpa, Miji and Sartung, on the East and Southeast by the Aka, and on the West by the Sherdukpen.

Van Driem (2007) estimates the population of the Buguns to be around 800. Other population counts of the Bugun vary. The 1991 Census of India reports a population of 1,046, while the 2002 census report shows a population count of 1,384. The Ministry of Tribal Affairs has enlisted the Buguns as one of the major communities in the Scheduled Tribes list in 1950.

2

The Bugun community witnessed two major cholera outbreaks in the 1950’s–1960’s, which decimated the community. Villages like Singchung have about 250–300 native speakers of Bugun, comprising 20% of the local population. The rest of the 80% population is made up of other local tribes, as well as Nepalis and other peoples from the rest of India. The Nepali population in Singchung is particularly large. Most of the Nepalis work as labourers in farms and fields belonging to the Buguns. The customary law of the indigenous tribes of Arnuachal Pradesh is a saviour. It protects the natives from losing their identity. But the impact of the non-native languages on the Bugun language makes the survival of the language a difficult one.

This paper is an account of my field trips to the Bugun villages of Singchung, Wanghoo, New Kaspi, Ramaling and a few others. I shall primarily talk about the community, their language and culture and give a description of the phonemic system of the language. As I am at the initial stage of this language documentation project my paper will focus on my findings of the first two field trips.

Causatives in Boro Prafulla Basumatary Gauhati University

Boro (Bodo), a Tibeto-Burman language spoken by the Boros of Assam, has causative constructions as one of its morpho-syntactic aspects. The purpose of this paper is to examine how morphological processes are applied to the formation of lexical and syntactic causatives in Boro. Irrespective of their transitive and intransitive functions, morphological and lexical causatives will be analyzed in relation to their corresponding verb classes.

In Boro the causative forms are the results of causative prefixation and suffixation. Some of the prefixes that are used in the formation of causatives have been shown below. One of the most productive causative prefixes is sw-, which turns an intransitive verb into a transitive one. For example (1a)-(1b): (1a) gotho-a gab-dwng. boy-SU cry-CONT ‘The boy is crying.’ (1b) bipha-ja gotho-khou sw-gab-dwng. Father-SU child-OBJ CAUS-cry-CONT ‘The father is making the child cry.’ Another productive causative prefix is khw-. This prefix also turns an intransitive verb into a transitive one. For example (2a)–(2b): (2a) mansi-a gwglwi-bai. man-SU fall-PERF ‘The man fell down.’ (2b) bima-ja gotho-khou khw-khlwi-bai. Mother-SU child-OBJ CAUS-fall-PERF ‘The mother has made the child fall.’ There are several other prefixes in the language that help in the formation of causatives, such as phu-sung ‘make something short’, phw-lao ‘make something long’, etc. These will also be treated in the paper.

3

The causative suffix -hw is freely used with any sort of verb, whether transitive or intransitive. This causative affix is highly productive in the language. For example (3a)–(3b): (3a) dwi-a gwdou-bai. water-SU boil-PERF ‘The water has boiled.’ (3b) ai-ja dwi-khou gwdou-hw-bai. mother-SU water-OBJ boil-CAUS-PERF ‘(My) mother has caused the water boil.’ In some cases, -hw amounts to a transitivizer, turning an intransitive verb into a transitive one. For example, mau-glung ‘shake back and forth’ (intransitive) > mau-glung-hw ‘cause something to shake back and forth’ (transitive).

There are other syntactic processes which are productive in the construction of causatives in the language. For example (4)–(6):

(4) bipha-ja phisazla-khou mansi-khalam-bai. father-SU son-OBJ man-CAUS-PERF ‘The father made his son become a (good) man.’ (5) zaduala-ja mansi-khou mauzi za-hw-bai. magician-SU man-OBJ cat be-CAUS-PERF ‘The magician has turned the man a cat.’ (6) bima-ja phisa-khou mwsa-zwng or-za-hw-bai. mother-SU baby-OBJ tiger-by bite-allow-CAUS-PERF ‘The mother has allowed her baby bitten by the tiger.’ An attempt will be made in this paper to analyse all the possible causative formations in the language.

Agreement in Thadou George Bedell

Payap University Thadou, also called Kuki, is a Kuki-Chin language primarily spoken in Manipur State, India and adjoining regions of India and Myanmar. According to Ethnologue (Lewis 2009), the total speaking population is 231,200. Like other Kuki-Chin languages, Thadou exhibits a system of agreement marking between verbs and (some of) their arguments. The categories relevant to Kuki-Chin agreement are shown in (a) with the corresponding Thadou pronouns. (a) first (exclusive) first (inclusive) second third singular kei, keima nang, nangma ama dual keini eini nangni amani pural keiho eiho nangho amaho Thadou pronouns differ for four persons and three numbers.

4

Verbal agreement in Thadou is restricted to subjects, and further restricted to person; there appears to be no agreement marking for objects or number. The Thadou particles which mark person agreement with the subject of a verb are shown in (b). (b) first (exclusive) first (inclusive) second third ka- i-, ei- na- a- They also function as possessive agreement markers with nouns. Examples of Thadou agreement are (c) and (d). (c) keiman twija baptaiz kachansahbou ahi; (Mt. 3:11) ‘I will baptize you with water’ (d) nanghon mi suhkhel nangaidamuva ahile, (Mt. 5:14) ‘if you forgive people’s wrongs’ In (c) the particle ka- marks agreement of the verb baptaiz chansah ‘baptize’ with its subject keima ‘I’ and in (d) na- similarly marks agreement of ngaidam ‘forgive’ with nangho ‘you (plural)’. The paper will explore the details of agreement in Thadou in comparison with that in some other Kuki-Chin languages. Most of the examples will be taken from a Thadou translation of the Gospel according to Matthew.

Are classifiers redundant? The discourse functions of classifiers in Assamese Gautam Borah

Tezpur University To the ordinary human perception things are either objects or masses. Objects are things that tend to exist in the infinite space by occupying a limited part thereof (e.g., a dog). A mass, by contrast, has the tendency to exist by occupying rather the whole of space (e.g., water). An object has thus a “shape”, or what Quine has called an “inbuilt structure”; a mass, without such a structure, is shapeless. Now, counting is repeating a shape, i.e. some finite space, where the number of repetition is encoded in the numeral. Thus, an object is numerically quantifiable while a mass is not, for infinity cannot be repeated. As linguistic counterparts of things, nouns thus subsume either a set of homogeneous objects, or a homogeneous mass. A typical numeral classifier is used with a noun that subsumes objects, not a mass; and in indicating the natural shape of these objects, a classifier not only classifies but also individuates the noun, i.e. it obliquely states that the associated noun is numerically quantifiable. Thus, in Assamese, for instance, which is a classifier language, tiniKHAN kitap ‘threeKHAN book’ ‘three books’ is grammatical, but *tini kitap is not, for the classifier KHAN, which categorizes flat and broad objects, is missing in the phrase to classify and thereby individuate the noun for numerical quantification. However, in clear contrast to classifier languages, nouns in languages such as Basque, Georgian, Hungarian, Oromo, Tsou directly go with numerals in their bare form. The following example from Tsou is from Rijkhoff (2002: 38). (1) iuso’o kueai

two.ART car ‘two cars’

These languages seem to indicate that classifiers, which classify and individuate nouns and thereby facilitate numeral quantification, are rather redundant: a noun like kueai ‘car’, for

5

example, itself indicates that what it refers to is an object and that therefore it is numerically quantifiable.

But when studied in the larger syntactic environment of discourse it becomes clear that classifiers are not redundant: they classify and individuate not just to facilitate counting (or reference, or possession) at the phrasal level, but also serve as anaphor to the noun that is just classified and individuated. Again, usually only an individuated noun can be made pragmatically salient or topically continuous in discourse. If so, one could expect classifiers to have important role in discourse. Thus, for instance, in (2) below, the noun buri ‘old woman’ is classified, for it has an important role to play in the subsequent discourse: (2) edin e-ZANI māgani buri xihatar

one.day one-CLF beggar old.woman their ghar-oloi sāul māgiboloi āhisil. house-DAT alms beg.INF come.PAST ‘One day an old beggar-woman came to their house to beg for alms.’

On the other hand, the noun sāul ‘alms’ and ghar ‘house’ in the example are left unclassified, for neither of them, unlike buri, is going to be developed in the subsequent discourse. Again, a proper name, since it is inherently fully individuated, is usually not individuated by a classifier. In (3), both nouns, i.e. bāndar ‘monkey’ and xiāl ‘fox’ are classified because they are the principal participants in the narrative: (3) e-TĀ bāndar āru e-TĀ xiāl āsil.

one-CLF monkey and one-CLF fox live.PAST ‘Once there lived a monkey and a fox.’

But in the subsequent mentions (e.g. in (4)) they are found unclassified. This is because they are now conceived as “talking animals”, i.e., they are now personified: (4) tāke dekhi bāndar-e xiāl-ak kole…

that seeing monkey-NOM fox-ACC say.PAST ‘Seeing that Mr. Monkey said to Mr. Fox…’

Thus, by classifying or leaving a noun un-classified, the speaker manipulates meaning in discourse. In this paper we make an attempt a detailed study of the discourse functions of classifiers in Assamese.

Reduplication in Boro Krishna Boro

Gauhati University

One of the prominent morphological features of Boro is reduplication. It is a very productive morphological process in the language. Almost anything can be reduplicated in this language. It is restricted neither to a particular word class nor to a particular semantic domain.

Both total and partial reduplication are found in Boro; for example (respectively), siri siri ‘silently’ and manswi sanswi ‘man and other things’. In the case of partial reduplication, both initial partial reduplication and final partial reduplication are found; for example, hazw hala ‘mountain or hill’ and akham sakham ‘rice’. However, among the two, final partial reduplication is more productive. In fact, the former is restricted to a few fixed expressions such as hazw hala.

6

In general, only one of the components of a reduplicated word is independently meaningful. For example, in the word hazw hala, hazw means mountain or hill, whereas hala does not mean anything. This is particularly true of partially reduplicated words. In case of totally reduplicated words, however, both the components may be either meaningful or meaningless. For example, in siri siri, the word siri means ‘silent’. In khwrw khwrw, on the other hand, none of the two components is independently meaningful; they are meaningful only as long as they are together.

Reduplication has a wide range of function in Boro. One of its important functions is to derive adverbs from verbs. For example, gele ‘play’ is a verb, whereas its reduplicated version functions as adverb. Consider the following examples (1)–(2):

(1) gotho-a gele-bai dwngw.

child-SUB play-PROG be.PRES ‘The child is playing.’

(2) i manswi-a gele gele mau-bai dwngw. that man-SUB casually work-PERF be.PRES ‘That man is working casually.’

Another important function is to add a sense of ‘plurality of things’ or ‘repetition of actions’ to the reduplicated word. In case of nouns, reduplication adds a sense of ‘and other related things’. For example, the noun akham means rice, and its reduplicated version akham sakham means rice and other related things like dishes. Consider the following example (3): (3) mauzi-a akham sakham za-zub-bai.

cat-SUB rice and others eat-end-PERF ‘The cat has eaten the rice and all other things.’

When an adjective is reduplicated it adds a sense of plurality to the noun which it modifies. For example, gaham manswi means ‘good man’, whereas gaham gaham manswi means ‘good people’. Similarly, a reduplicated adverb indicates, depending upon the nature of the action, that the verb which it modifies denotes a repetitive action. For example, the verb bu ‘beat’, when modified by a non-reduplicated adverb such as gwsa ‘hard’, denotes a single instance of beating, but when modified by a reduplicated adverb such as gwsa gwsa ‘hard’, it denotes a repetitive act of beating.

This paper will provide a detailed description of the formation and functions of reduplicated words in Boro.

Adjectives in Boro Bihung Brahma

Central Institute of Technology, Kokrajhar, Assam Adjectives in Boro are essentially nominal forms, which either precede or follow a noun. Like nouns in Boro, adjectives can inflect for case, and in many cases they are synchronically derived via nominalization. Although the majority of adjective stems in Boro are bimorphemic, there are several simple adjective stems with no internal structure. Mono-, di- and trisyllabic stems occur. Among these, disyllables are the most common, as in: mwzaŋ ‘good’

7

gazri ‘bad’ khathi ‘nearby’ gwzan ‘far’ Many unanalyzable nouns function freely as unmarked modifiers, and by that criterion belong to the adjective category: phisa ‘child; small’ seŋra ‘young (of a male); young boy’ sikhla ‘young (of a female); young girl’ bwrai ‘old (male); old man’ burwi ‘old (female); old woman’ However, most of these cannot occur in a comparative construction, and thus are perhaps better analyzed as being more nominal than adjectival.

The largest number of Boro adjectives consist of a stem bearing a prefix gV-. The commonest form of this prefix is gw-: gw-zam ‘old, ancient’ gw-dan ‘new’ gw-zou ‘tall’ In many words the vowel of this prefix harmonizes with the stem, especially when it contains a high vowel: ga-ham ‘good’ gi-zi ‘old; torn’ gi-si ‘wet’ gu-suŋ ‘short’ However, this process is not predictable. There are several examples in which the prefix has a vowel other than /w/ which does not match that of the stem: go-tha ‘solid’ gu-war ‘wide’ ga-zri ‘bad’ Most gV- adjectives correspond to unprefixed verb stems: be gotho-a gwswm. this boy-SUBJ dark ‘This boy is dark (in complexion).’ be gotho-a swm-bai. this boy-SUBJ darken-PERF ‘This boy got dark (in complexion).’

8

The phonology and syntax of Mizo language Sangluaii Chhangte

Govt. College, Serchhip, Mizoram In this paper, I will present my observations, experience and knowledge about the phonology and syntax of the Mizo language. The Mizo language, which has its own distinctive nature, is related to the Tibeto-Chinese languages.

The Mizo language and the Mizo tribe itself is said to be descended from the Tibeto-Chinese group of people, who are sub-divided into several groups. The Mizos are placed together with Burman, Meitei (Manipuris), Naga, Kochin, Lolo, Tibetan, etc.

The Mizo alphabet was framed and designed by British missionaries in the year 1894, and the Mizo language was for the first time reduced to writing by the British Missionaries F. W. Savidge and J. H. Lorrain using a Hunterian system of orthography. The Hunterian orthography was perhaps never adopted elsewhere, nor is it traceable anywhere in the world.

The Mizo language (Mizo Tawng) is spoken by the Mizo people who live within Mizoram. There are different kinds of dialects in Mizoram, such as Lushai, Hmar, Lakher, Lai, Paihte, etc., as there are different clans in Mizoram amongst which the Lushai dialect is the main medium of communication and lingua franca within Mizoram.

The first remarkable feature of Mizo language is that it is a tonal language, in which variation in tone and intonation pattern can change the meanings of words and utterances. In Mizo language, there are four distinct tones, namely, Rising, Falling, Mid-low and High. A change of tone indicates a change in the meaning of the word. In the Mizo System of writing, however, tonal markers are rarely used, and the meaning of a word is generally made out from the context of use. Therefore, words with the same spelling can in fact have different meanings according to their tones.

The phonological tones of Mizo language will be illustrated in the paper. Some syntactic differences of Mizo language from other languages such as English, Hindi, etc., will also be shown.

Going here and there in Mizo: Moving in time and space Lalnunthangi Chhangte

Mizo is a language spoken in the state of Mizoram. It belongs to the Mizo-Kuki-Chin branch of Tibeto-Burman languages. This paper examines the function of directionals in the language, especially with respect to discourse. Mizo indicates direction and location by two methods, both of which can occur concurrently.

Location, with respect to the speaker and addressee is indicated by two determiners in the NP (1). (1) He mi hi

this man this ‘this man (here)’

There are six possible pairs: Hei hi (this here), Cu cu (that, out of sight), So so (that away from speaker and addressee), Kha kha (that away from speaker), Khu khu (that down there), and Khi khi (that up there).

Direction and motion are indicated by directionals in the VP. These occur between the subject agreement prefix and the head verb (2).

9

(2) Ka-han-en. I-up-look ‘I went up there and looked.’ or ‘I looked up.’

There are four possible directionals: low and ron (both towards speaker), va (away from speaker) han (up away from speaker), zuk (down away from speaker).

These sets of directionals also have a separate function in the context of discourse. For example, determiners in the NP are used to introduce participants in the discourse. Similarly, directionals in the VP can express the speaker’s attitude toward the event, sometimes taking on exclamatory functions. The paper will examine in detail the features mentioned above with examples from spoken texts.

Reciprocal constructions in Asamiya Runima Chowdhary Gauhati University

This study is a descriptive analysis of the various strategies Asamiya employs to express reciprocity. These strategies used for marking reciprocity may be broadly classifiable into three types: syntactic, morphological and lexical.

The syntactic strategy involves the use of a closed set of different types of pronominal reciprocals functioning as anaphors. The pronominal reciprocal is a bipartite structure, morphologically consisting of two different case-marked components with distinct semantic roles. The reciprocals of this type requiring plural antecedents can be inflected for various cases. The study reveals that case-copying is a distinctive feature of an anaphoric reciprocal construction, whereby its initial component invariantly copies or repeats the case of the given antecedent.

The morphological strategy involves derivation of reciprocals of various categories either by amalgamation of the processes of affixation and reduplication or by compounding. The former process productively derives reciprocal nouns, verbs, adjectives and adverbs from non-reciprocal roots, whereas, the latter being a relatively less productive process derives compound nouns only. The language in its lexical strategy employs various categories of inherently reciprocal lexemes. The lexical reciprocals include non-derived nouns, verbs, adjectives, adverbs and postpositions which are needed to be kept in the mental lexicon of the native speakers.

Number building, numeration systems in Pnar, War, Khasi and Lyngngam and some of their Austroasiatic features

Anne Daladier LACITO-CNRS

Pnar, War, Khasi and Lyngngam (PWKL) cardinal numerals subdivide into two subgroups: Pnar, Khasi, Lyngngam on the one hand and War on the other one. This subdivision parallels other lexical and typological features and probably corresponds to a settlement in two main subgroups of the speakers of these languages together with different Pnar communities in Meghalaya; see Daladier (forthcoming).

10

These two cardinal systems appear to be composite, as is usual in both MK and Munda systems; see Jenner (1976) and Zide (1978). They show traces of different numeration bases and borrowings from Indo-Aryan and from Thai. In addition to its cardinal numbers, PWKL still uses counting units with their own names. These names are found in different Munda and MK cardinal systems with shifts in their decimal cardinal values.

Counting units depend on the goods counted and involve different numeration bases. For example ʃi həli ‘one unit of four pieces’ is used for counting citrus, betel nuts or eggs, and all other things which must be counted in a quadrennial basis. Various multiples of four are used as units for specific goods. Betel nuts, chillies, dry fish and paan leaves, as opposed to eggs, oranges and lemons, can also be counted in tā (War) / kti (Pnar, Khasi, Lyngngam) ‘(both) hands’ that is in a decimal base. ta/ti ‘hand’ for the decimal cardinal ‘five’ appears in some Munda and MK systems as a remnant trace of a quinary numeration base. Those three goods can also be counted in a vigesimal basis, sometimes with different units. For example, a unit of twenty pieces for paan leaves is ʃi bdi and for fresh fish it is ʃi kuri; however, ʃi kuri of paan leaves amounts to 20 ksɛp; that is, 1600 paan leaves. The notion of number used in PWKL counting units is different from the notion of number in Hindu-Arabic or Chinese cardinal systems. A PWKL numeral used as a counting unit does not have an absolute cardinal value, but rather has a cardinal value relative to what is counted.

PWKL units of large cardinality specific to classes of objects (with their peculiar baskets) prove useful in dealing with large numbers without written technical devices.

Quadrennial, quinary, decimal and vigesimal numeration basis can be used alone or can combine in PWKL units. These numeration bases are also found as remnant traces in other MK and in Munda cardinal systems.

Some cardinal numbers are expressed as subtractions and are specific to PWKL, but subtraction devices as building blocks for number systems are also found in Munda; see Zide (1978). There are very interesting similarities between South Munda and PWKL numerals.

Many cardinals of other MK and Munda languages can be related to counting units still used in trade activities in PWKL. There are secondary shifts in the cardinal values of these unit names in the decimal systems of AA languages, which may have several causes. One cause may be confusion between different numeration basis used to express the cardinal value of a numeral. Another cause may be reductions occurring in subtraction or addition devices from or to a unit. For example, PWKL has retained an opposition between ‘one’ as a cardinal and ‘one’ as a unit, using for the second an element related to the cardinal ‘ten’ in many other MK languages. There are shifts in the cardinal values of several AA numerals, like ti/ta ‘hand’ for five or sometimes for ten. Some IA and Thai cardinal numerals have also been borrowed in MK and in Munda with shifts in their cardinal values. A way to account for these shifts would be to assume that at first they have been integrated in the unit systems of AA languages with AA own numeration bases, and then integrated into Hindu Arabic decimal cardinal systems together with other AA units.

One wonders whether such units might be traced back to trade accounts between Ahom and Pnar kings.

Genitive case marking in the subject: Some evidence from Assamese Gahin Ch. Das

Abhayapuri College, Assam

Typologically, Assamese is a Nominative-Accusative SOV language, and its prototypical subject takes the Nominative marker. Verbs agree with the subject in person only. But in certain expressions we find a different kind of NP in the subject slot which is in the Genitive case. The verb may be either in the finite or nonfinite form and it does not agree with the subject.

11

(1) mor bhuk lagise. 1.SG.GEN hunger feels ‘I am feeling hungry.’

(2) mor zor uthise. 1.SG.GEN fever is.rising ‘I am coming down with fever.’

(3) mor (tomaloi) bhoe lage.

1.SG.GEN you fear feels ‘I feel scared (of you).’

This paper presents examples of Subject NPs from Assamese which are in the Genitive case, and discusses their semantic, syntactic and behavioral properties in relation to different predicate types.

The role of reduplication in the process of Assamese word-formation Dipak Das and Khirapada Dutta

Madhabdev College, Assam Reduplication is one of the most natural processes of word-formation. Yet as Abbi points out, “although reduplication is one of the most important processes of forming words, most of the grammar books ignore the phenomenon, considering it rather trivial” (Abbi 2001: 161). Defining the process, Bloomfield says, “reduplication is an affix that consists of repeating part of the underlying form” (Bloomfield, 1933: 218). Abbi defines reduplication as, “words formed by duplicating syllables or by duplicating a single word (phonological word), partially or completely” (Abbi, 2001: 162).

Reduplication plays an important role in Assamese word-formation. Assamese, a member of the Indo-Aryan language family, uses reduplication both syntactically as well as in a variety of other functions. For example, abundant uses of lexical, semantic, and expressive reduplications, as well as echo-formation, are found. Instances of class-changing and class-maintaining as well as partial and complete reduplications are also found in the Assamese language.

In fact, reduplication is an inseparable part of Assamese word-formation. Many words of this language are formed via this process. However, until now, no systematic study has been done in this field. Though mention of various grammatical items of this type is found in certain traditional Assamese grammar books, yet no attempt has been made to study this potent word-formation process as an independent item of grammar.

North-East India is an important linguistic area which is home to many different language families. Each language has played a role in shaping its neighbouring language(s). Therefore, this type of study can give us a better understanding of the formation and development of particular languages. This paper will make an attempt to study the role of reduplication in the process of Assamese word-formation.

12

Language, culture and biodiversity: A critique of the formal legal systems of North-East India

Joy Dasgupta and Sarit Chaudhuri Greater Himalayan Bio-Cultural Society, Shillong, Meghalaya

International discourse about the need to protect biodiversity-based traditional knowledge is epitomized by a host of international conventions and treaties, most notably Article 8(j) of the Convention on Biological Diversity. The overarching discourse at the international level has been to recognise the importance of traditional knowledge systems and the contributions made by traditional/indigenous people, and their importance as major actors in the conservation of biodiversity is highlighted in the article 8(j) of the CBD. However in the context of the north-east India, there has been a very paternalistic acknowledgement of the importance of traditional knowledge by emerging state institutions in the field of biodiversity management. The emerging formal legal structures have tended not to take into consideration the complex linkages between the transmission of oral knowledge and the languages of the region. The discourse has instead been one of looking at language, culture and biodiversity as three distinct and discrete entities that are not intertwined. This paper argues that it is in fact critically important for formal legal systems to evolve a nuanced understanding of these interlinkages, acknowledge their interconnectedness and develop an incentive-based policy regime to ensure that all three elements are protected equally and are not looked at as discrete and disparate entieties. The approach should move beyond mere documentation of traditional knowledge systems to a more interconnected approach that acknowledges the equal importance of this triad, namely, language, culture and biodiversity.

Boro adverbial suffixes and the genesis of grammar Scott DeLancey

University of Oregon

Though its Tibeto-Burman heritage is clear in its lexicon and in a few unmistakable traces of ancient PTB morphology, Boro shares with its more distant Tibeto-Burman neighbors little more than typological and areal features of morphology and syntax. It has been suggested that this is because its most immediate ancestor, Proto-Bodo-Garo, ca. 2,000–2,500 BP, was a highly creolized lingua franca in the Brahmaputra Valley, and that the grammar of Boro and the other BG languages for the most part was formed at that stage and since.

All of the Bodo-Garo languages have an unusual category of verbal elements which we call “adverbial suffixes” (following Burling’s (2004) “adverbial affixes”). Examples of this category are khaŋ ‘to completion’ and hwi ‘at a distance’ in (1).

(1) bwrai-a bhwttw gi-khaŋ-hwi-bai

oldman-SU afraid-COMPLETE-DISTANT-PERF ‘The old man was overcome by fear (away off there).’ (Mirror)

These morphemes carry a wide range of meanings, from abstract, essentially aspectual

senses like phin ‘again’ and khaŋ ‘to completion’, to highly lexical senses like khao ‘into pieces along the linear axis’. We have no clear idea of how large the category might be, but Boro has well over a hundred of them, and possibly a great many more. These constructions look very much like verb serialization of the sort reported in other TB languages, especially in Lolo-Burmese. But while many of these suffixes correspond to synchronic verbs, many others, perhaps a majority, do not. Some of the most lexical are onomatopoeic in origin.

13

Neither we nor other linguists who have worked on closely-related languages (see especially Burling 2004: 139–153) have been able to find much internal grammatical structure to this system. The most lexical suffixes occur next to the lexical verb, but otherwise they occur in no fixed order; the three sentences (2a)–(2c), with three different orders of the adverbial suffixes thar ‘completely’, phin ‘again’, and lai ‘each other’, are all possible, with no evident difference in meaning. (2a) gau-zẃŋ-gau bu-thar-lai-phin-dwŋ.

self-with-self beat-completely-reciprocal-again-ASP ‘They are killing each other again.’

(2b) gau-zẃŋ-gau bu-thar-phin-lai-dwŋ. ‘idem.’

(2c) gau-zẃŋ-gau bu-lai-thar-phin-lai-dwŋ. ‘idem.’

Many of the adverbial suffixes can be sorted into broad (but very leaky) semantic categories, but these seem to have no reflection in grammar.

What we see here is the very beginnings of a grammatical system, or perhaps more than one, which in Boro has barely begun to coalesce out of its roots in what must have been a system of very free verb serialization, with little or no formal marking.

A sociolinguistic study of proper names: A case study of the Meiteis of Manipur Ch. Sarajubula Devi

NCERT, Shillong This paper is an attempt to visualize Meitei personal names through sociolinguistic perspectives. The focus of the paper is: first, to classify Meitei personal names by origin with an idea of defining a rough historical period from which the name groups emerged; second, to give a classification of Meiteilon personal names; and, lastly, to study the trend of Meitei socio-cultural revivalism as it is reflected in Meitei personal names.

Meitei personal names are not given at random. Rather, a name is chosen with knowledge of what that name implies (i.e., the semantics of the name concerned). Analysis of Meitei personal names reveals that most of the proper names used in Meitei society are Hindi/Sanskrit-based names; the next category is personal names of Meiteilon origin; a few are of English origin, and some are of mixed origin. Meiteilon personal names of English and other origins are usually extravagant names. In my finding, Meiteilon personal names are subclassifiable into six groups as: order of birth, quality, abstract, physical appearance, ability-based, and the last one is the name of the flowers being extended to personal names. Though many personal names remain to be of Hindi-Sanskrit origin, there is a strong tendency to adopt personal names of Meitei origin, probably as a resultant of the strong revivalism trend visible in the Manipuri society. Many people are consciously trying to use or adopt a Meitei personal name as a nickname. The trend is clearly visible in the choice of nicknames by younger folks.

The data used for this analysis is based on a collection of a randomly selected list of personal names numbering around 770 which were supplied by seven teacher trainees from Manipur in the North Eastern Regional Language Centre, Guwahati-28. They

14

represent different parts of Manipur. Each of them were requested to provide at least one hundred and ten common personal names each, of which five hundred have been utilized after canceling the repeated ones.

Negation in Nyishi: A fieldwork report and analysis Moumita Dey

Jawaharlal Nehru University Nyishi, a Tibeto-Burman language and the mother tongue of the Nyishi tribes of Arunachal Pradesh, is an isolating and agglutinating language. Nyishi, the cover term used to refer to all the Tani languages, means ‘the land of people – the humans’. With a rich stock of expressives, and rich case-inflectional affixations, the word-order varies between being SOV and SVO, with varying varieties. This paper aims to present a field report and analyse the same in terms of negation in Nyishi, the data for which was taken from informants, some of which were students and some senior citizens, visiting Delhi, for academic and commercial purposes.

In Nyishi, the primary negative morpheme is mα, which is used whenever a negative sense is implied. It is used in sentential negation, as a negative suffix, and as a negative particle. Though the negative morpheme mα alone carries on its monosyllabic shoulders the responsibility of negating any given element or context as per the needs, the only other negative word found is αpαlαi for ‘without’. The word order of the language mandates that the negative particle mα is placed post-verbally – after the verbal morpheme – as a suffix and carries with it the sense of mode – namely, whether the sentence or the clause refers to the realis mode or the irrealis mode in the grammar. To negate a given verb such as yubnΛm, ‘to sleep’, the infinitive morpheme nΛm is replaced by negative mα to form the sense ‘not sleeping’ as in (1):

(1) yub-nΛm → yub-mα-do sleep-INFTV → sleep-NEG.REALIS-PROG ‘to sleep’ → ‘not sleeping’

As constituent negation, the negative word mα is positioned after the given constituent, the noun or the noun phrase. It is seen that the negative morpheme mα always follows its scope, the constituent whose sense is to be negated, in the language. As we move mα across a given sentence, the informants tend to interpret the sentence differently, giving different meanings, in turn changing the scope of negation in the grammar. mα becomes rəm in the environment of an irrealis mode and into yo in that of the imperative mood. But elsewhere, retains its original form as in (2): (2) mα → rəm /__ V + irrealis mode + NEG mα → yo /__ V + imperative mood + NEG

mα → mα /__ V + elsewhere + NEG

In case of hortative moods, a repeated structure is used in the sentence, that is, kuj tuj, where kuj begins the sentence and tuj ends the same. The negative counterpart of such hortative statements replaces the end-word tuj, with buj, which is preceeded by the negative suffix -mα. Thus, kuj buj indicates the negative hortative mood as in (3): (3) kuj αsα-kəm əmə-mə-mα buj.

HORT other-ACC lie-CAUS-NEG HORT.NEG ‘Lets not lie to others.’

15

However, in conditional negative sentences, we find that conditional morphemes always follow the negative morpheme mα, probably because the conditional morpheme acts as a conjunction and joins the two clauses. However, the negative members of the shifter classes are different and individually distinct. Negative pronouns such as those below are positioned before the verb in a sentence:

hiyənə ‘nobody’ hiyə ‘none’ hoggo ‘nothing’ huglo ‘nowhere’

However, along with these words, we find the obligatory presence of the word kəm ‘at all’, together meaning ‘not at all’, as in (4): (4) αlo hiyə kəm do-mα.

there none at.all be-NEG ‘There is none out there.’

The semantic diversity of mα in Nyishi is varied and wide and serves to basically negate almost every affirmative instance in the grammar. But their use and application at various places differ, leading to diverse forms and types of the negative morpheme mα, for it is used not only as a negative word, but also as a negative particle, a negative suffix, etc. It is this aspect of negation, which this paper aims to work out in the report, following their analysis, based on the data collected.

Adjectival nominalization patterns in Mising Karen Grunow-Hårsta and Sarat Doley

University of Wisconwin at Milwaukee and Lakhimpur College

Tibeto-Burman languages frequently do not support an independent category of adjective. Though these languages may have a very small class of “true” adjectives, the native TB pattern is to treat concepts that would be rendered as adjectives in other languages (for example in most Indo-Aryan languages) as nouns or, more commonly, as stative verbs. In TB, such “adjectival verbs” are typically nominalized; i.e. the same affix is used to form agent, patient and event nominals, and it may also be used to form adjective (relative) clauses. True adjectives, both native and borrowed, typically do not undergo derivational (nominalization) processes. This paper investigates how, and to what degree, the derivation of Mising adjectivals aligns with these generalizations as they are attested in other TB languages; for example in Chantyal (Tamangic), Magar, Kham and Newari (Himalayish), Mongsen Ao (Kuki-Chin-Naga) and Galo (Tani). It also discusses the ramifications of divergence from the “typical” TB pattern of nominalization in Mising.

A preliminary overview of East Bodish Gwendolyn Hyslop

University of Oregon The East Bodish branch of Tibeto-Burman has been discussed primarily by George van Driem in several publications (e.g. van Driem 1998), but little is in fact known about it. Though East Bodish languages are primarily centered in Bhutan, some are also spoken in Tibet and Arunachal

16

Pradesh. In this paper I present data from several Mainstream East Bodish (MEB) languages of Bhutan and Arunachal Pradesh in a preliminary overview, reconstructing forms for Proto-MEB when possible.

The MEB languages are separated from their Central Bodish cousins (i.e. the Tibetan dialects) by a number of phonological, lexical, morphological, and syntactic differences. A brief comparison of personal pronouns and numerals suffices to show this: e.g. WT <khyod> vs. Proto-MEB *i ‘2.SG’ and WT <bzhi> vs. Proto-MEB *ble ‘four’.

Within MEB there is at least one obvious sub-grouping based on both lexical and phonological criteria. As has been suggested in van Driem (2007), Dzala and Dakpa appear to form their own subgroup within East Bodish. An example of the lexical differences which distinguish this group is the word for ‘sun’, which is ne in Kurtöp and Bumthap, ni in Khengkha, and nam in Chali, but plang in Dakpa and prang in Dzala. Dakpa and Dzala also regularly have /e/ where the other languages have /a/, and /i/ where the other languages have /e/. For example, ‘horse’ is ta in Kurtöp, Bumthap, Chali, and Khengkha but is te in Dzala and Dakpa, and the word for ‘four’ in Kurtöp and Bumthap is ble and in Chali bre, but in Dakpa it is bli and in Dzala bri.

Another fairly obvious grouping is that of Kheng, Bumthap and Kurtöp, which have been argued to form a subgroup of their own (e.g. van Driem 1998). However, within this group, there are phonological reasons to keep each of them separate. For example, where Kurtöp and Bumthap have an aspirated velar-labiovelar initial /khw-/, Khengkha has innovated a voiceless bilabial fricative /ɸ/. Furthermore, Kurtöp has innovated a series of retroflex stops from velar plus rhotic clusters still present in Bumthap and Khengka (e.g. Kurtöp ʈong but Bumthap/Khengkha krong ‘village’).

Further sub-grouping of the East Bodish languages remains problematic. Tentatively, I can show that Khengkha, Kurtöp, Bumthap and Chali are separated from Dzala, Dakpa and Phobjip based on the sound change l > j (e.g. Dakpa/Dzala lɛŋe and Phobjip laŋ vs. Kurtöp/Khengkha/Bumthap/Chali jaŋa ‘five’). However, the Phobjip word for ‘ten’ khepche shares more in common with the other MEB languages than with Dzala and Dakpa (ci and ciŋnəi, respectively). Third person singular pronouns suggest that Chali and Kurtöp may be closer to one another (khi and khit, respectively) than to Dakpa and Dzala, which have the forms ba and be, respectively. The Bumthap and Khengkha form gon appears to be a recent innovation (present still in Kurtöp as ‘friend’).

Our understanding of the East Bodish languages is in its infancy, and the looming question ‘to what extent has Classical Tibetan (the liturgical language for almost all East Bodish speakers) influenced the East Bodish languages?’ remains the most pressing problem in placing East Bodish languages within the larger Tibeto-Burman family, as many items which could be used for sub-groupings are suspiciously (and usually unsystematically) similar to Tibetan. Nonetheless, this paper is a first step in sorting out the relationship amongst the East Bodish languages, and eventually, the relationship of the East Bodish languages to the rest of Tibeto-Burman.

Grammaticalization of the nominative marker =a into a discourse marker in Meche

Kazuyuki KIRYU Mimasaka University, Japan

Meche is a western variety of the Bodo language spoken in the Jhapa District of Nepal and West Bengal. This paper deals with a case of grammaticalization of a case marker into a discourse marker in Meche. The data are taken from the Jalthal Dialect of Nepal.

Bhattachariya (1977) discusses two markers for nominative and accusative in Bodo. Zero-marking on nouns is regarded as a general nominative and accusative, while -a is a definite

17

nominative and -khow is a definite accusative. The explicit case markers are definite in function, and Meche has exactly the same patterns of case marking. However, it is often the case with Meche that an object noun is doubly marked with =a and =khəu (= Bhattachariya’s khow), as in (1).

(1) (bərai=ya) burəi=ya=khəu phəsi-dəŋ. old.man=NOM old.woman=DM=ACC support-PROG ‘(The old man) supported his wife.’ Furthermore, some object nouns take the putative nominative marker =a without the accusative marker =khəu. (2) gənəi burəi=ya wa ̉ţ=na chu-mat-dəŋ nə. then old.woman=NOM fire-DM light-actually-PROG HS ‘The old woman was indeed making a fire.’ Based on further data, I will discuss that the marker =a that appears with objects functions as a discourse marker (DM), marking a prominent participant in the discourse, regardless of definiteness. Discourse prominence has to do with contrastive focus, topicality and discourse cohesion. Also, I assume that the DM was originally the nominative marker, and that it later grammaticalized into a discourse marker.

Koch dialects of North East India Alexander Kondakov

SIL International Koch is a tribe of the mongoloid stock found mostly in the North-East Indian states of Meghalaya and Assam, parts of West Bengal and in northern Bangladesh. Till the present there has been a dearth of research on Koch. This paper describes a socio-linguistic survey of Koch varieties which was carried out from 2006 to 2007.

The very name ‘Koch’ has for long been a term of some ambiguity often denoting quite different things to different people. The Koch discussed in this paper are those who have preserved their tribal dialects, animistic religion and non-Hindu cultures and traditions. They divide themselves into the nine ethnolinguistic groups: Tintekiya, Wanang, Harigaya (Songga), Margan, Chapra, Satpari, Sankar, Koch-Rabha (Kocha) and Banai. Dasgaya appears to be an alternate name of both Margan and Chapra groups. Koch-Rabha or Kocha of Lower Assam are linguistically Koch; however, about 50–60 years ago they officially joined the Rabha tribe in order to get some facilities from the state.

The Koch groups are endogamous and their members do not intermarry with members any group other than their own. Under each of the aforesaid endogamous groups there are small exogamous clans called nikini. The Koch are matrilineal, although there is no hard rule like among the Garos and the Khasis that in all cases a man must go to live in his in-law’s house. Among the Koch groups, there are very few intermarriages.

The Koch are agriculturists. The vast majority of them still live in villages and lead a largely traditional mode of life. The Koch religion is traditionally animism; however, they have been to a large extent Hinduised.

Since there has been much confusion regarding the identity of the Koch, it is hard to determine the precise number of their population. Various sources suggest various figures: 14,000 persons in 1971 (the Census of India); 16,682 in 1981 (ibid.); 23,000 in

18

2005 (Gordon 2005); S. N. Koch stated in a private conversation that the Koch population of Garo Hills could not exceed 30,000 persons.

The language the Koch speak goes by the same name. Linguistically, Koch has been classified in the following way: Sino-Tibetan > Tibeto-Burman > Jingpho-Konyak-Bodo > Konyak-Bodo-Garo > Bodo-Garo > Koch. Each ethnolinguistic group of Koch is thought to speak its own dialect. However, the speakers of only five groups, namely Wanang, Harigaya, Tintekiya, Margan, and Koch-Rabha have preserved their Tibeto-Burman mother tongue. The other groups, such as Satpari, Sankar and the most of Chapra have abandoned their native language and speak an Indo-Aryan tongue, a mixture of Assamese and Bangla.

The relationships between the five Koch speech varieties are rather complex. They represent a dialect chain that stretches out from Koch-Rabha in the north to Tintekiya Koch in the south. These speech varieties are found in relation to one another as such:

Koch-Rabha → Wanang → Harigaya → Margan (Dasgaya) → Tintekiya The adjacent dialects show greater lexical similarity, while those at the ends of the chain

show lesser similarity. In addition, there is interesting evidence of sound change across dialects.

An investigation of Karbi tones

Linda Konnerth University of Oregon

Karbi is a Tibeto-Burman language spoken mostly in the Karbi Anglong district of Assam, Northeast India. In his Karbi grammar, Grüßner (1978: 18–22) describes a three-way tone distinction between low-mid-high; all three tones are analyzed phonologically as level tones. Phonetically, Grüßner notices typical patterns of slightly falling contours in low tone open syllables, and slightly rising contours in high tone open syllables.

The more interesting case is the mid tone. First, according to Grüßner, it has a restricted distribution as it does not occur on obstruent-final syllables (i.e., syllables with final unreleased voiceless stops and a few instances of final /s/ in borrowings), whereas both other tones occur on any type of syllable. Second, phrase-finally, this tone is characterized by a following glottal stop.

This paper compares Grüßner’s analysis with data on monosyllables recently gathered from a native Karbi speaker during fieldwork in Karbi Anglong. The essential three-way distinction can be replicated by a few minimal triplets. However, the data pose several problems for the further analysis of the Karbi tonal system. Although I could also find that one member of the minimal triplets features an audible glottal constriction in isolated repetitions, it appears not to be the one with the medium pitch level, but with the highest frequency – although this might actually not be consistent. Another interesting finding is that this speaker exhibits a case of vowel lowering from /i/ to /e/, which might stem from an interaction with the tonal system.

By addressing some of the phonetic features that interact with F0, such as vowel duration, vowel quality, and voice quality, this paper tries to give an account of the pitch patterns found in the collected data.

19

Documenting Paite Bornini Lahiri, Atanu Saha and Ritesh Kumar

Jawaharlal Nehru University

Paite is a Tibeto-Burman language mainly spoken in Churachandpur District, Khuga Valley, and Copur Bazar in Manipur. It is also spoken in some parts of Mizoram, Tripura and Assam. Outside India, some variety of Paite is spoken in Myanmar also. With a population of just around 50,000 (according to Census 2001) and the restricted use of language in home domain (English being mostly used for educational and other purposes), the UNESCO Atlas of Endangered Languages has classified it as an “unsafe” language. Considering this fact we started documenting the language as part of our M.Phil. coursework on ‘Language Documentation’.

Language documentation is seen as “a lasting, multipurpose record of a language” (Himmelmann et al. 2006). According to Austin, language documentation should be “diverse”, “large”, “ongoing, distribuited, and opportunistic”, “transparent”, “preservable and portable” and “ethical” (Himmelmann et al. 2006). We have tried to document Paite keeping in mind these benchmarks.

The present documentation constitutes of mainly three aspects – documenting the sounds and all the phonetic, phonemic and prosodic features of the language by annotating as many words as possible, documenting the lexicon and presenting it in the form of a multimedia, multilingual dictionary and documenting the ethnolinguistic and cultural aspects of the community as well.

The sounds of the language (phonemes along with the tones) are annotated in Praat and a phonetic database of the language is being built which we are planning to send to the MBROLA project for use in the speech synthesis. The dictionary is made using Toolbox by SIL and it is presented in the form of an interactive dictionary through LexiquePro. The ethnolinguistic and cultural documentation consists of documenting terms and concepts from a wide and diverse areas that included marriage, religion, kinship terminology and metalinguistic knowledge of the language experts.

In the present paper we want to share our experience of documenting a language in a field away from the field and what we have managed to document. The documentation, as of now is diverse (but not so much as it ought to have been) but not large. It is portable and ethical. And it is definitely ongoing.

Presentation on the Tangshang (Tangsa) Naga in Myanmar Ahkhi Lisu

This presentation will describe the language research I have done among the Tangshang (Tangsa) Naga in Myanmar. I will present some background information about the Naga groups living in Myanmar, including their location, classification, and living conditions. I will show and describe the results of my lexicostatistic analysis of wordlists I have gathered from approximately 30 Naga varieties in Myanmar. In addition, I will discuss the implications of these wordlist comparisons in terms of potential intelligibility and recommendations for language development efforts among these Naga groups.

20

Towards deciphering the linguistic content of an age-old Dimasa narrative

Monali Longmailai North Eastern Hill University, Shillong

Dimasa is a Bodo-Garo language of the greater Tibeto-Burman family. The speakers of Dimasa are found in Assam and Nagaland. Their origin can be faintly traceable from an age-old oral narrative which is the only etymological record available today. This reminiscent verse has been passed on through generations of the Kachhari clan Zaunthai, a wise expert. The verse narrates the migration of the Dimasas and other Kachharis from their faraway origins to the present land; that is, the northeastern region of India that dates back to around 3,000 BCE.

This oral narrative appears in written form along with its literal meaning in Chapter 1 of the book “The Queens of Cachar or Herambo and The History of the Kachchari” (2007) authored by the eminent Dimasa historian Nalinindra Kumar Barman. Except for historians, no linguist has ever made an attempt to decipher the narrative which is full of complex morphology and syntax still retaining its proto word forms, no matter the distortion, besides its intrinsic poetic language. Some of the word forms are still in use today and they give clues not just about the origin of the Dimasas but also provide immense scope for the historical reconstruction of the Proto-Dimasa language. This paper aims to analyse the traditional narrative of Dimasa from a morpho-syntactic perspective. It will look into the etymology of the words, the structure of simple and complex words and affixes, and the possible rules used for word-formation in the narrative. In addition to this, it will also decipher the semantics of the verse.

Exploring the deictic system of Tagin Language Kepor Mara and Sudhanshu Shekhar

Jawaharlal Nehru University Tagin is the name for both a language and the people who speak it. Tagin people inhabit the Upper Subansiri district of Arunachal Pradesh. This language has not attracted the eyeballs of many linguists so far. Slowly but steadily this language is going into the dungeon of language imperilment. We could not find out any full linguistic description of this language. In this paper, we will endeavour to capture one of the interesting aspects of Tagin and that is its deictic system. Deixis is derived from the Greek word deickto. By deixis we mean, all clues encoded by the language and given to us in grammatical form in order to localize a speech event and its participants (speaker and hearer) and narrated objects in space and time.

The categories of deixis are personal, temporal, spatial, discourse and social deixis. Deixis concerns the ways in which languages encode features of the context of that utterance or speech event and thus also concern the ways in which interpretation of utterance depend on that context of utterance.

The many facets of deixis are so pervasive in natural languages, so deeply grammaticalized, that they could be easily thought of as essential part of semantics. However, deixis is a part of pragmatics because it directly concerns the relationship between the structure of a language and context in which it is used. Deixis is a particular kind of reference which depends on the time and place of utterance and upon the speaker’s and an addressee’s role in the utterance itself. The deictic system of Tagin shows many interesting facts and triggers many facts to be explored.

The deictic system in this language is a harbinger of the fact that language is a complex system of codes which is guided by both external and internal phenomenon. There are many kinds of deixis that are functional in Tagin. We will try to incorporate numerous interesting

21

functions of various kinds of deixis in this paper. Tagin has a plethora of time deixis. Tagin people have eleven different words to denote the different times of day. Porok logu is the expression which is used to denote the period very early in the morning when the cock creates the sound COCKOO! COCKOO! Aru-kəmci is used to denote period early in the morning. There are four directions in this language. They are: up, down, left and right. In this language North, which equals up, is primarily given in terms of the altitude of the reference. The direction also changes with the movement of the speaker. In the case of rivers, Tagin speakers have four directions, but the words used for directions are different from the words used for direction in general. The pronominal system also showcases the social hierarchy of this language.

The kinship system and vocabulary of Milang Yankee Modi

Jawaharlal Nehru University

Milang is a Tani tribe mainly found in the Upper Siang district of Arunachal Pradesh. Traditionally aligned with the Eastern Tani languages (Sun 1993), Milang at the same time shows a much greater number of distinctive differences from other Tani languages than is the Eastern Tani norm. Milang is the only Tani language which is completely mutually unintelligible with all other Tani languages. As part of a larger cultural-linguistic documentation and description project which aims both to characterize several distinctive features of Milang and to compare them with parallel features in other Tani languages, the present study aims to characterize the kinship system of modern Milang and its linguistic representation.

The marking of noun phrases: Some observations on the languages of North East India

Stephen Morey Research Centre for Linguistic Typology

Traditional grammatical descriptions of languages spoken in North East India often rely on the classification and terminology used for the classical languages: Latin and Sanskrit. Thus one often reads about “Nominative”, “Accusative”, “Dative” and so on, relating to languages as widely divergent as those of the Tibeto-Burman, Indo-Aryan and Tai families.

One of the features of the Latin and Sanskrit systems is that they are systematic: every noun must be marked by one or other of these cases, and at least some of these markers are clearly syntactic in nature: for example, the subject of a sentence appears in the nominative case. These systems form paradigms that are learned by students as the declensions of Latin and Sanskrit/Pali. These cases are marked both on the heads of noun phrases (the noun), as well on the various elements that modify those nouns, such as adjectives and demonstratives.

In the much more diverse languages found in the North East of India, however, we often see a very different kind of NP marking, with very different features:

• marking is not always obligatory, and therefore does not form a paradigm • marking may depend on the animacy of the noun • marking may combine several of the categories that are separated in Sanskrit and

Latin (particularly the use of a single marker to convey both PATIENT and RECIPIENT, marked by Accusative and Dative respectively in Latin and Sanskrit)

• marking may cover a whole noun phrase, rather than marking the head noun • modifiers like adjectives and demonstratives receive no special marking

22

In this paper we will survey the marking of noun phrases in a number of languages spoken in the North East. We will demonstrate that the terminology and categories of the classical languages are often unhelpful for the study of these languages, and suggest the use of alternative appropriate terminology such as “agentive”, “anti-agentive” or “non-agentive” rather than the classical terms.

Stage level vs. individual level predicates and the choice of copulas in Assamese Diganta Kumar Nath

Tezpur University

In typological studies, it has been observed that different languages have different copula systems. Unlike English and Hindi, which use one copula, some Indian languages use more than one copula. Languages like Marathi and Malayalam use two distinct copulas. Oriya uses four different copulas, namely aT-, ach-, tha:-, and he-. Bangla also has four copulas. In this paper, we are going to argue that Assamese makes use of four copulas in Assamese, namely a:s, ha, tha:k and Ø (null). All of them, however, do not behave in the same way. Each of them has it’s own features and characteristics. For example, a:s functions both as an auxiliary and as a copular main verb, indicating locative and existential use. The locative copula a:s takes the negative form na:i ‘NEG.be’. One peculiarity is that the copula a:s takes personal agreement but its negative counterpart na:i does not take any agreement marker. One peculiarity of the ha copula is that, unlike the a:s copula, it is never used in past tense. However, ha has two different realizations hOl and hOise, which are used in past tense. A peculiarity shared by these two copulas is that they are never used in negative sentences. In the literature, they are called positive polarity copulas. Again, the copula tha:k ‘stay’ is existential in nature. In the literature, copula sentences are sometimes distinguished as either ‘Equative vs. Existential’ or ‘Equative vs. Predicative’. However, these binary distinctions are not adequate to account for the four copulas which are available in Assamese. In fact, Mahapatra (2002) shows that neither of them is helpful in predicting copula choice in Indian languages. In this paper, we are going to examine the four copulas in Assamese, using the ‘Individual-Level’ vs. ‘Stage-Level’ predicate distinction proposed by Carlson (1977) to account for temporary vs. permanent time reference in the context of verbal predicates. In a sentence like (1), both a:se and tha:ke are existential copulas. However the use of tha:ke gives a generic reading of the sentence. It has a permanent time reference compared to that of a:se. (1) Pukhurit pa:ni a:se/tha:ke. Pond.locative water exist ‘There is water in the pond’ Thus, in this paper we will first argue that there are four copulas in Assamese. This would resolve the typological divide between Assamese on the one hand and Oriya and Bangla on the other hand. Then the reminder of the paper will attempt to look at the four copulas in Assamese, using the ‘Individual-Level’ vs. ‘Stage-Level’ predicate distinction proposed by Carlson (1977). I hope this would give us an insight into the choice of copulas in Assamese.

23

A case study of linguistic variation among the Monpa and other related varieties spoken in Arunachal Pradesh

Liahey Ngwazah NLCI, Bangalore

This paper summarizes a sociolinguistic research conducted among several linguistic varieties (3 varieties of Monpa, Sherdukpen, Aka, Chug, Lish, Bugun and Miji) of Western parts of Arunachal Pradesh. The fieldwork was carried out at various times between November 2003 and August 2004 by Binny Abraham, Kara Sako, Isapdaile Zeliang and Elina Kinny. The research investigated linguistic variations, intelligibility, language use, and language attitudes, using tools such as wordlists, Recorded Text Testing, and questionnaires. The following were the conclusions:

1. There are significant variations between 3 Monpa varieties. It is very unlikely that one language development project would meet the needs of this tribe.

2. The other 6 languages too are distinct from each other, with the highest divergence being

that of Aka. Within the Aka, there are two separate dialects that are not related to each other.

Two types of verbal predicates in Tiddim Chin OTSUKA, Kosei

University of Tokyo, Japan In this presentation, I will discuss nominal predicates and verbal predicates that are found in VPs in Tiddim Chin, a Tibeto-Burman language spoken in northwestern Myanmar and North-East India. I deal with colloquial Tiddim Chin, which is spoken on the side of the Union of Myanmar (Tedim/Tonzang Township, Chin State).

There are two types of predicates (VPs) in Tiddim Chin which Henderson (1965) did not mention in his previous studies, i.e. nominal predicates and verbal predicates. Both types occur not only in subordinate clauses, but also in main clauses. We easily tend to overlook this feature as most verbal clauses are often nominalized without any nominalizer in colloquial Tiddim Chin; however, the two types are so distinct from each other that it is essential to describe this aspect of colloquial Tiddim Chin grammar. Looking through the texts that I have collected in Myanmar, I will show how both of the predicates behave syntactically and pragmatically in colloquial Tiddim Chin. Some of the distinctive features of verbal predicates and nominal predicates in VPs are shown below. [1] Verbal predicates (VPs) (1-a) Pronominal proclitics are postposed to verbal predicates. (1-b) A copular verb hì: cannot be followed by a verbal predicate. (1-c) An enclitic =kéi is applied as a negative marker. (1-d) Irrealis mood is indicated by postposing an irrealis pronominal proclitics. [2] Nominal predicates (VPs) (2-a) Pronominal proclitics are never postposed to nominal predicates. (2-b) A copular verb hì: can be followed by a nominal predicate. (2-c) An enclitic =lòu is applied as a negative marker. (2-d) Irrealis mood is indicated by postposing a purposive marker =dí:ŋ.

24

The Siyom River Valley: An essay on intra-subgroup convergence in Tibeto-Burman

Mark W. Post Research Centre for Linguistic Typology, LaTrobe University

The most salient feature of Sun’s (1993) Tani stammbaum is a division into two major branches: Western Tani and Eastern Tani. By and large, the Western/Eastern division is well-supported on phonological, lexical and grammatical grounds. However, there are several “problem cases”; namely, languages which do not fall neatly into either branch, but instead exhibit certain features from both branches. Galo is such a language, which was “tentatively” classified as a “peripheral member” of Western Tani by Sun; in Sun’s view, Galo had a basically Western Tani core, but nonetheless exhibited several Eastern Tani traits which complicated the effort at classification. In this paper, I will first argue that Sun was basically correct, and that Galo is a genetically Western Tani language. However, a prolonged period of contact along the West > East-flowing Siyom River with the neighbouring Minyong (Adi) – an Eastern Tani tribe – has led to a certain amount of cultural and linguistic convergence, which has complicated the genetic picture considerably. The remainder of the paper will attempt to characterize this convergence. Some of the issues to be considered will include: the diffusion of cloth-weaving technology and vocabulary, the diffusion of dispute-resolution structures and concepts, the diffusion of certain phonological changes (such as *ɕ- > h- rather than s- and *rj- > r- or j- rather than lj-), borrowing of lexical items (in both directions), convergence of grammatical structures, and the historical and contemporary meanings and uses of the tribal/linguistic label “Adi”.

Question formations in Asamiya (Assamese) Nibedita Sarma

Gauhati University

Interrogation in Asamiya (Assamese) has some special properties. In this study, explanation of the various types of questions in Assamese, and the various grammatical inflections of k-words is attempted from the perspective of linguistic research. Examples from both general (also known as closed/polar) and special (also known as open/content) quetsions are taken into consideration for this explanation. Normally, general questions in Assamese are formed via suffixation of a question clitic /-ne/ to the finite verb or insertion of a question particle /neki/ after it. These two formatives /-ne/ and /neki/ are in complementary distribution. Tag questions are formed by suffixing /-ne/ to the invariant affirmative tag verb /hɔe/ ‘is’ or the negative tag verb /nɔhɔe/ ‘is not’. Alternative questions are formed by the use of question clitic /-ne/ between the two possible choices of answers. Open questions are constructed by replacing the relevant constituent with a k-initial question word or “k-word”. There is no movement of the question word in Assamese. These k-words are inflected for various cases and in the pronominal position these words may be suffixed with classifiers, numbers and genders.

Tiwa Tones in Monosyllables Priyankoo Sarmah

Hankuk University of Foreign Studies (Korea) Tiwa, also known as Lalung, is a Tibeto-Burman language of the Bodo-Garo subfamily, and has 26,480 speakers in Assam state in India. Not much is known about the phonology of the Tiwa

25

language. In this work, I investigate the tones of the Tiwa language with the help of instrumental acoustic measures supported by statistical tests in order to determine the types of tones that exist in the language. I also examine the acoustic characteristics that can help native speakers to distinguish the tones in Tiwa. Furthermore, I investigate how tones operate in non-derived monosyllabic words in the language.

Tiwa, according to Joseph and Burling (2001; 2007) has two contrasting tones, which are high level and falling. They postulate that there can be either a high-level or a falling tone on a monosyllabic word. Speculating that either syllable in disyllables in Tiwa is capable of hosting either one of the two lexical tones, Joseph and Burling (2001; 2007) claim that disyllabic words can have four possibilities to host lexical tones: i) the first syllable has a high tone, ii) the first syllable has a falling tone, iii) the second syllable has a high tone or iv) the second syllable has a falling tone. In case of iii) and iv), the first syllables are assigned a neutral tone. In case of i) and ii), they claim that the tone tends to spread to the syllable on the right.

The goal to investigate the Tiwa tone system here is three-fold: a. the number and inventory of tones in Tiwa b. tone assignment in monosyllables c. determining the acoustic cue for identifying tones in Tiwa Acoustic analysis of the tones in this language revealed that like some other Bodo-Garo languages Tiwa also has a two-way tonal contrast. However, unlike other Bodo-Garo languages, Tiwa exhibits only a contour tone system with a rising and a falling tone. Additionally it was also revealed that the two Tiwa tones differ from each other both in terms of direction of the contour (i.e. rising or falling) and average pitch (F0). In addition to that it was also found that contrary to the claims in previous studies the duration of tone is not dependent on the type of tone. Hence, it was concluded that only pitch or F0 is the primary cue in the identification of tones by Tiwa speakers.

Sibilants and glottal spirants across Assamese dialects Shobha Satyanath University of Delhi

It is widely acknowledged that the Middle Indic palatals, both voiced and voiceless, shifted to alveolar sibilants in Assamese. In the same way, the pre-existing sibilants shifted to glottal spirants (Grierson 1903; Kakati 1941; Goswami and Tamuli 2003). The first shift, affecting the palatals [c/j] to [s/z], is widespread across the North-Eastern region, cutting across languages. The shift from sibilant to glottal spirants is, however, relatively less common. It is more productive in the varieties related to Assamese or the Indo-Aryan varieties spoken in Assam and elsewhere within the North-East. The present study explores the shift of “s(h)ibilants” to “hibilants” (or, following Grierson, “s(h)ibboleths” and “hibboleths” (Grierson 1903:V(I), p. 201)) across dialects of Assamese located within the present day Assam.

While the shift from palatals (both voiced and voiceless) to sibilants is almost categorical in Assamese, it is to be noted that not all sibilants become glottal spirants. The shift is further selective as it affects essentially the pre-existing sibilant phonemes, thereby discriminating the original sibilant phonemes and the sibilants arising out of the pre-existing palatals. The shift is also subject to a number of constraints both phonological and morpho-syntactic. This study proposes to investigate the shift of sibilants into glottal spirants. It intends to provide a comprehensive description of the linguistic constraints that trigger and block such a shift. Although such shifts are generally referred to as lenition, I refrain myself from using any such term.

26

Some of the shifts, particularly the spirantization of palatals, are often attributed to contact with the Tibeto-Burman languages that are spoken alongside Assamese in Assam (Grierson 1903; Kakati 1941). This explanation, though probable, needs to be revisited. Similarly, it is an open question as to whether the shift of sibilants into glottal spirants might be a contribution of Tai Ahom contact. In contrast, the attestation of similar phenomenon across Assamese dialect regions may certainly be attributed to the forces of contact and diffusion.

The specific objectives of the study are listed below: 1. a detailed description of the shift of sibilants into glottal spirants and its relationship with

the other related shifts. For instance, whether or not, and in what ways the palatals, the sibilants (both original and derived) and the glottal spirants (original and derived) are related to each other and to the shifts under discussion. In other words, whether it is at all possible to link the three shifts as part of a single cumulative progressive change such as [c > s > h]

2. constraints on the shift of voiceless sibilants into glottal spirants 3. whether the shift under discussion forms a single unified phenomenon across Assamese

dialect regions 4. the motivations for such a shift – to what extent the various shifts can be attributed to

non-Indo-Aryan contact and to what extent these are the results of internal developments 5. a comparison of the constraints across relevant speech varieties to estimate the role of

diffusion on the constraints. The study is based on two kinds of data (i) the available historical texts (ii) modern rendering of the earlier texts and similar data. The analysis is descriptive, interpretive and quantitative when needed.

Elements of indefiniteness: the -ba forms of Assamese Jyotiprakash Tamuli Gauhati University

This paper takes a beginning look at a set of words in Assamese that end in -ba. What all these k-initial words have in common is that all of them express a sense of indefiniteness: kiba 'something', kɛnɛba ‘somehow’ kʊnʊba ‘someone’ etc. They also have counterparts in the language without the -ba element: ki ‘what’, kɛnɛ ‘how’, kʊn ‘who’. The absence of similar indefinite expressions in other Indo-Aryan languages such as Bengali or Hindi, and the attested presence of the -ba element in a Tibeto-Burman language such as Garo makes the -ba forms of Assamese an interesting object of study.

The investigation will be based on a section of data from the untagged CIIL-EMILLE Assamese corpus of written Assamese.

Comparison of selected Tangsa varieties in Changlang District of Arunachal Pradesh

Mathew Thomas Annamalai University

Tangsa is a cover term that encompasses a group of tribes in Arunachal Pradesh who speak linguistic varieties belonging to the Tibeto-Burman family of languages. So far, an appropriate linguistic designation of “language” or “dialect” has not been assigned to the linguistic varieties

27

spoken by Tangsa communities. The present study is limited to exploring the linguistic situation of varieties spoken by Tangsa communities in the Changlang district of Arunachal Pradesh. According to the 1981 Census records, there are thirty-two Tangsa tribes within India (Singh 1995). Among these tribes, a few (Mosang, Kimsing, Jugli) have tried to produce literature in their respective varieties. None of the literatures produced in any of these varieties could meet the literature needs of all of the Tangsa communities. It was in this context that the investigator deeply felt the need to undertake a study on the Tangsa varieties and their level of relatedness to one another, and thereby gaining clarity on the linguistic status of these Tangsa varieties in Changlang District. This study would empower the Tangsa communities to decide on whether they can continue with the existing literature or produce new literature.

Initially, a word list comprising 176 lexical items was collected from all the available Tangsa varieties (25) for a lexical similarity analysis as described in Blair (1990). According to him, different dialect areas can be distinguished by the process of lexical similarity comparison and dialect intelligibility testing. Frequent inter-tribal contact and intermarriage exists between Tangsa communities; therefore, the results of the lexical similarity analysis and intelligibility tests did not adequately sort the relatedness between Tangsa varieties. Accordingly, the researcher opted to compare the phonology, morphology, syntax and meta-linguistic data of a few selected varieties.

The present study would like to propose that Tangsa varieties need to be viewed as different linguistic varieties. The existing similarities between Tangsa varieties can be accounted-for by the convergence that has taken place due to frequent inter tribal contact and the sociolinguistic network existing among Tangsa communities. Data will presented basically from two Tangsa varieties in Changlang; Mosang and Haidley with some examples from a few other Tangsa varieties too.